Prediction of chemical bond formation using efficient CUDA based HPC framework

High Performance Computing (HPC) is now days a keyword which talks about efficient and cost effective computing systems or frameworks. Applications with huge computations are dependent upon the hardware support and the parallel algorithms designed and executing on this hardware. Research in the domain of parallel computing obtain the parallel algorithm which efficiently use the now days available multi-core hardware such as NVIDIA's GPUs. This paper work contributes to design, development and analysis of such HPC framework using GPUs which would support the execution of prediction algorithms. As an application, we are predicting the possibility of formation of chemical bond formation among atoms/molecules and their possible compounds. We present the results of execution of these algorithms on sequential and parallel platforms which show the concrete improvement in the execution time and the efficiency.